Transfer Pipe: Acoustic Conductor; Passive Probe?

How about an A^#₈/B^b₈ tuning fork? (If they go that high.)

Electronic stethoscope technology might be a good approach.

I don't know what to suggest but http://www.globalspec.com/ is often a good source to look at.

You really don't give us much to go on here for a recomendation. I realize that you intend to do relative amplitude measurements but it would help greatly if you could give people an idea how loud you expect your acoustic signal will be. The information you provide implies that you expect the transducer must be immersed in the oil but this will certainly add a tremendous amount of turbulance itself that will certainly include acoustic noise in your narrow bandwidth. I would try a contact microphone or acceleromator attached to the outside of the pipe first. You could then go with a high gain band pass amplifier that would reduce your interferring frequencies.

Guys, Thanks for the feedback. I will check on the signal level and get back to you. I believe the signal level with be in the range of -60 dbm to 0 dbm, but will try to narrow that down. I can't put a probe into the oil flow. The inner wall of the pipe and manifold needs to be smooth bore.

I could try an accelerometer on the surface of the manifold, but I think the mass of the manifold, at 1.8" thick, would dampen the signal to below -130dbm, and make it undetectable.

Where are you injecting the frequency and what is the overall goal? We would need to see the specs for the injection probe. What is the pipe diameter? Why was that particular frequency chosen? 7.2 kHz seems very specific to a purpose.

Without pipe diameter (to give oil velocity) and the pressure in the pipe (1.8" could stand a lot of pressure!) which no-one can guess, a safe device will be difficult to suggest. Try a pressure transducer or an external accelerometer first.

The pipe is 10" schedule 40 carbon steel pipe. The manifold is carbon steel and is 1.8" thick. Pressure is 175 psi. I cannot access the pipe with instrumentation, only the manifold. The 7.2 Khz signal is being kinetically induced upstream in the system. I need to monitor the signal at the manifold, both frequency and amplitude. We will be making parametric changes in the process upstream, so I need to monitor the result at the manifold.

I am thinking an anelastic polymer rod with a low mass metallic needle centered in the rod might do what I need. I can bore the manifold and bond in the rod. If the needle can conduct the 7.2 Khz signal from the hot oil to the surface of the manifold, I should be able to read it with an accelerometer.

As redfred suggested, I can build a bandpass amplifier and notch out extraneous bearing and motor signals. I just need to get the 7.2 Khz signal out to a point where I can read it without disturbing the oil flow.

Just some suggestions for tests. I am not an acoustics man (originally telecomms), but some of the techniques are similar to microwave radio "plumbing". A sketch of test rig proposed is below. It may help to identify unhelpful acoustic effects - like resonances around 7.2 kHz in your "plug".

1. Polymer plug seems good idea. I do not think a metal rod will help much. Ideally, you have good acoustic matching to transfer energy out of the pipe into your receiver. Oil [density 0.9] to steel [density 8.5] is going to cause a big reflection at the interface - what is reflected will not go to your pick-up. Polymer, with similar density to oil, looks better.
2. Extend polymer plug well out of manifold - low thermal conductivity will take your pick-up away from high temperature giving cheaper pickup & less thermal drift. Also makes attachment of sound pipe feasible. You can probably super-glue small accelerometers onto the end of the plug.
3. Experiment with a test rig drawn below - pipe may also allow you to do pressure test on your polymer insert before use!
4. I have used loudspeakers scavenged from old radios to do resonance checks on electronic PCBs, mountings and housings. I cut off the cone, leaving the support spider and attach a thin wire to the moving part at the middle using a wire spider and epoxy resin. Attaching wire to PCB or chassis and driving loudspeaker from a basic 600 ohm 5 volt rms sine audio signal generator was adequate (with a cheap electret microphone - easily interfaced to most PCs with microphone input - which provide the DC power needed to the microphone input) to check there were no bad resonances in the vibration test range. Sticking the microphone on the board or holding it close to board was effective (mostly, I could hear the bad resonances without microphone!). N.B. Computers with a sound card have a 1Watt or so 32 ohm output - and you can get signal generator programmes with sine output at adjustable frequency - your FFT programme may have one as an accessory.
5. As shown in sketch, you can attach a thin diaphragm to the oil pipe to launch a wave into oil. A small loudspeaker in air in a plastic pipe over the end of the oil pipe might work also (PTFE tape diaphragm?). Or bury a moving coil earphone element from a cheap 32 ohm earphone headset in the oil, wiring through an oil-tight gland. Assuming your oil is not very electrically conducting... ordinary mineral oils are used for transformer insulation - it's those with additives which conduct. Easy enough to tell - see if the earphone resistance changes when you immerse it in an oil sample.
6. 7200 Hz will propagate in a 2 cm diameter (or less) air pipe with plane waves and set up standing waves. Moving the microphone back and forth to find a place where the acoustic impedance suits the microphone (to give maximum pick-up) may help. The wavelength in air at N.T.P. for 7k2 Hz is about 4.75 cm, so a standing wave will go max to min over about 1.25 cm - should be possible to get an "organ pipe" resonance as an acoustic filter. N.B. the wavelength will be a lot less in a solid/liquid - higher velocity.
7. Putting insert in end of manifold is probably best, supposing 7.2 kHz is propagating along the manifold.
8. Since you know your frequency, "synchronous detector" electronics, with a phase-locked loop may prove useful. Averaging with a long time-constant can extract signals totally buried in noise. But FFT is good at isolating distinct frequencies and probably easier for you to try.
   

Maybe, an acoustics man with a solution to hand will surface from the CR4 depths...

Best wishes,

67model

67M, Thanks, this is helpful. I actually did some RF work in my early days. RF Broadband amps, and Gigahertz range cavity resonators, and wave guide stuff. You have jogged my memory a bit. I think a plastic rod built to a 'tuned' length and diameter may pass the 7.2K signal from 300 degree oil up to 70 degree air as you suggest.

So, now I need a plastics guy to chime in on this thread. I assume the density of the plastic rod will determine the propagation delay, and hence the 'wavelength'. Once I have the material and wavelength, my machinists can build some rods to test out.

Anyone?

CR4 has a "Chemical & Material Science" forum - see the right hand side menus on CR4 pages. A post there may bring help quicker.

Some more ideas out of the microwave/audio box. Make plastic plug tapered (in steel screw-in housing), ideally exponential like a horn loudspeaker or antenna - may help with mechanical stress resisting that 175 psi! 2) use phonograph pickup needle/cartridge on side of straight sided bare plastic specimen as standing wave probe to get Standing Wave Ratio SWR as indicator of matching/getting weakest echos to distort your measurement. N.B. sensitive axis transverse to groove in vinyl disc 3) Silicone rubbers have high mechanical hysteresis - hence damping; could work as tapered "waveguide" terminator load - like the wooden wedges in microwave waveguide.